Nowhere to hide: This drone can track targets through the forest.

A team of researchers from Austria recently developed an AI-powered drone that can track moving objects through dense foliage. Too good to escape into the woods and live on land if machines oppose us.

by team research paper:

Detecting and tracking a target moving through leaves is difficult (and often impossible) in general aerial imagery or video, and is practically feasible through image integration…

This discovery, along with the implementation of early drone-operated camera arrays for parallel synthetic aperture aerial imaging, allows us to present the first results of tracking people moving through dense forest. Objects other than humans (such as vehicles or animals) can be detected and tracked in the same way.

This can affect many applications such as search and rescue, surveillance, border control, and wildlife observation.

Front: There are a myriad of ways an interested person can track a moving object through dense foliage, including FLIR and other thermo-optic systems, but this offers a new AI-powered ability to track color change through wrinkling, or occlusion.

neural newsletter header

Humanoid greeting

Subscribe now to get a weekly summary of your favorite AI Stories.

Teaching AI to figure out what’s going on in an image when a significant amount of relevant information is hidden is one of the biggest challenges in the AI ​​world. In this case, the researchers have developed a system that can track moving occluded objects in real time using a technique called “color anomaly detection”. remote sensing journal.

background: Using drone deployments, the team created a lightweight 1D camera array that captures overhead images of defoliated areas. Conventional aerial photography techniques use color anomaly detection to scan an image, looking for pixel clusters that do not match their natural environment.

However, the team’s contribution was to develop a combination of hardware and machine learning that combined existing technology with a new aerial camera system to create something that could identify color anomalies and track them through occluded areas.

Image of a drone with a camera array next to the failure of an array component.
Credit: Nathan, et al.
Image of a drone with a camera array next to the failure of an array component.

quick time: This can provide immediate and potentially huge benefits to search and rescue efforts. The relatively low computing requirements and power overhead make this solution a ready-to-service solution that can be shipped relatively quickly to remote locations around the world. I can see this saving the lives of off-course hikers or plane crash survivors trapped in remote areas.

In addition, when used in conjunction with GPS tagging and other tracking efforts, conservation efforts can benefit greatly.

However, something like this could turn a Predator drone into something that, for example, can track a target through a closed area (such as a congested city block or jungle) even if its connection to a human controller is lost or communications are disrupted. It is also clear that there is

Leave a Comment